Design Platform for Sustainable Catalysis with Radicals: Electrochemical Activation of Cp2TiCl2 for Catalysis Unveiled

Abstract The combination of synthesis, rotating ring‐disk electrode (RRDE) and cyclic voltammetry (CV) measurements, and computational investigations with the aid of DFT methods shows how a thiourea, a squaramide, and a bissulfonamide as additives affect the EqCr equilibrium of Cp2TiCl2. We have, for the first time, provided quantitative data for the EqCr equilibrium and have determined the stoichiometry of adduct formation of [Cp2Ti(III)Cl2]−, [Cp2Ti(III)Cl] and [Cp2Ti(IV)Cl2] and the additives. By studying the structures of the complexes formed by DFT methods, we have established the Gibbs energies and enthalpies of complex formation as well as the adduct structures. The results not only demonstrate the correctness of our use of the EqCr equilibrium as predictor for sustainable catalysis. They are also a design platform for the development of novel additives in particular for enantioselective catalysis.


General Information
All reactions involving air-or moisture sensitive compounds were prepared in oven dried glassware under inert atmosphere (Ar) using standard Schlenk and vacuum line technique. All chemicals were purchased from abcr GmbH, TCI GmbH, Alfa Aesar, Fluka, Acros, Flourochem or Sigma Aldrich and were used without further purification. Solvents used in manipulations under inert atmosphere were either dried and deoxygenated by distillation (THF over Na/K-alloy) before use or purified inside a M-Braun MB-SPS-800 solvent purification system and used after degasification. The NMR analysis was carried out on a Bruker DPX 300 MHz, Bruker 400 MHz or Bruker 500 MHz spectrometer. 1 H and 13 C NMR chemical shifts were specified in ppm and calibrated by using the residual undeuterated solvent as internal reference [ 1 H NMR: CHCl 3 (7.26 ppm), C 6 HD 5 (7.16 ppm), DMSO-d 5 (2.50 ppm), CHD 2 CN (1.94 ppm), THF-d 7 (1.72 ppm, 3.58 ppm); 13 C NMR: CDCl 3 (77.0 ppm), C 6 D 6 (128.0 ppm), DMSO-d 6 (39.5 ppm), CD 3 CN (118.3 ppm), THF-d 8 (25.3 ppm, 67.2 ppm)]. High resolution mass spectra analysis was measured on a Thermoquest MAT 95 CL or Thermo Fisher Scientific LTQ Orbitrap XL instrument. IR spectra were obtained on an ATR-IR Spectrometer Thermo Electron Nicolet TM 380 instrument as neat film. CHNS analysis was measured on an Elementar Analysesysteme varioMICRO instrument. Silica gel (230-400 mesh) supplied by Merck and Macherey-Nagel or neutral aluminum oxide 90 supplied by Merck was used as stationary phase for column chromatography.

Synthesis of Substrates
The substrate 1 was prepared according to the literature procedure. The additives L1 [S1] and L2 [S2] were synthesized following the cited literature. Analytical data of the corresponding products can also be found in these references. L3 was synthesized in analogy to the cited literature. [S3]

General Information
The chemicals tetrabutylammonium hexafluorophosphate (NBu 4 PF 6 ), tetrabutylammonium chloride (NBu 4 Cl) and AgNO 3 were purchased in electrochemical grade from Aldrich and stored in a glovebox under an inert atmosphere (Ar).
All cyclic voltammetry experiments were carried out in a glovebox and were performed by a 1400D Electrochemical Analyzer (CH-Instruments). A glassy carbon disk of 1 mm diameter was used as working electrode material. The surface of the working electrode was polished with diamond paste

General Procedure for Cyclic Voltammetry Experiments
The conducting salt NBu 4 PF 6 (0.775 g, 0.200 mmol) was dissolved in the freshly distilled solvent (10 mL) and a magnetic stir bar was added to an oven dried cyclic voltammetry cell. Background measurements were conducted to reduce coulomb currents in the analysis by subtraction from the CVs recorded with analyte. Titanocene dichloride (5 mg, 0.02 mmol) was added to the cyclic voltammetry cell and CVs were recorded. After this, the indicated amount of additive was added. The cyclic voltammetry experiment was carried out at different sweep rates (0.05 Vs -1 , 0.1 Vs -1 , 0.2 Vs -1 , 0.5 Vs -1 , 1 Vs -1 , 2 Vs -1 , 5 Vs -1 , 10 Vs -1 , 20 Vs -1 , 50 Vs -1 ) and the solution was stirred before measuring the respective sweep rate. At the end of the experiment a small amount of ferrocene (0.02 mmol) was added as an internal reference and the potential of the Fc + /Fc redox couple was recorded.

CVs of the Additives L1 -L3
Here, the CVs of L1 -L3 at 0.2 and 2 Vs -1 are depicted for reference:

Bulk Electrolysis
The divided bulk electrolysis cell was purchased from ALS which is distributed by C3 Prozess-und Analysentechnik GmbH in Germany. All controlled potential electrolysis experiments were carried out in a glovebox and were performed by a 1400D Electrochemical Analyzer (CH-Instruments). A glassy carbon mesh electrode was used as the cathode and a platinum wire immersed into a solution of CH 3 CN/NBu 4 PF 6 (0.2 M) and Cp 2 TiCl 2 (0.01 M) served as anode. The reference electrode is composed of a silver wire and a solution of NBu 4 PF 6 (0.1 M) and AgNO 3 (0.01 M) in CH 3 CN. Figure S1: Bulk Electrolysis Cell used for catalyst activation.

Controlled-Potential Electrolysis in THF:
The bulk electrolysis cell (100 mL) was equipped with a stirring bar and NBu 4 PF 6 (4.65 g, 12.0 mmol), Cp 2 TiCl 2 (149 mg, 0.60 mmol), additive (L1: 300 mg, 0.60 mmol, CPE1; L2: 342 mg, 0.60 mmol, CPE2; L3: 396 mg, 0.60 mmol, CPE3) and THF (60 mL) were added inside a glovebox. The electrodes were put inside the cell and connected to the potentiostat. The electrolysis was conducted at a controlled potential (E = -1.4 V), during the experiment the stirring was set to 260 rpm and the current flow was monitored. As an indicator the color change from red (Ti IV ) to green (Ti III ) was used and the electrochemical reduction was stopped when the current flow was below 1 mA. The electrolyzed solution was used as is for the arylation.

General Procedure for the Arylation of Epoxides (GP 1):
A pressure stable Schlenk flask was filled with a stirring bar and the epoxide (1.0 eq.). The solution (CPE 1 to CPE3) containing the electrochemically generated [Cp 2 Ti(III)Cl]-catalyst (0.01 M, 0.1 eq.) was added inside a glovebox. After that, the reaction mixture was stirred for the indicated time at the indicated temperature. The reaction mixture was allowed to cool down to ambient temperature, the solvent was evaporated under reduced pressure and the crude product was purified by column chromatography.
The data are in agreement with the literature.

Rotating Ring-Disk Electrode Experiments
Here, a more detailed discussion of the RRDE experiments is included. The discussion is started with measurements of the base electrolyte (2 mM Cp 2 TiCl 2 and 0.2 M NBu 4 PF 6 (TBAPF 6 ) in THF) and will continue with L1, L3 and L2 according to the increasing complexity of the systems.

General Remarks
In cyclic voltammetry [S6] mass transport is only given by diffusion. This is why detected currents for an electron transfer event run through a peak value at a potential, where the concentration of the analyte at the surface is close to zero and the flux of species towards the electrode is at maximum.
The following decrease of current density is a result of the diffusion layer extending into the bulk electrolyte, which leads to a decreasing concentration gradient. With a rotating electrode like a rotating disk or ring-disk electrode [S7] a continuous flow carries fresh bulk solution in vertical direction toward the electrode surface. However, close to the surface, the velocity of the electrolyte moving orthogonal to the electrode decreases to zero and changes into a radial flux due to the rotation of the electrode. Thus, a thin layer of electrolyte with a constant thickness is established in front of the electrode, through which the analyte has to diffuse. At the boundary to the bulk solution, the analyte concentration always equals the bulk concentration. As a consequence, a plateau rather than a peak is observed in a CV experiment at an RRDE once the concentration of the analyte at the surface of the electrode is close to zero. This is the diffusion limited current I lim . In a four-electrode set-up, with a disk and a ring electrode around it, the species formed in the original redox event at the disk or in a fast follow-up reaction are transported to the ring and can be detected. By adjusting the constant potential at the ring electrode, different species can be discerned.
This feature of the RRDE makes the method perfectly suitable for our aim to analyze the different species already observed in CV measurements. All measurements were conducted at a sweep rate of 25 mV s -1 with a reductive sweep followed by an oxidative back-sweep. Each measurement was performed at four rotation frequencies f (4, 9, 16 and 25 Hz). A thin-gap glassy carbon disk and ring (AFE7R9GCGC, Pine Research) with a disk surface area of A Disk = 0.247 cm² and a theoretical collection efficiency of N 0 = 0.37 was used as the working electrode. In order to put our later results into relation, we started with a measurement of Cp 2 TiCl 2 in 0.2 M NBu 4 PF 6 /THF, which is the base electrolyte. This system was already studied in depth in the Daasbjerg group. [S8] The ring potential was set to 0.05 and 0.66 V vs. Ag/Ag + . The voltammograms are shown in Figure S3. Note that the disk potential has been iR-corrected. The uncompensated resistance was determined via electrochemical impedance spectroscopy and was in the range of 500 Ω. Note that the iR-correction does not severely influence the evaluation of the equilibrium constants as the constants are evaluated from diffusion-limited currents and transfer ratios. S13 Figure S3. Disk current densities i D (a) and normalized ring currents I R /N 0 (b) of the RRDE measurements of a solution of 2 mM Cp 2 TiCl 2 in 0.2 M TBAPF 6 /THF with a ring potential of 0.05 V (I) and 0.66 V (II) at different rotation frequencies. The disk potential has been iR-corrected.
According to the E q C r mechanism, [S9]  Therefore, the ratio of ring and disk currents I R /I D is higher at 0.66 V than at 0.05 V as both Ti (III) species are detected at the higher potential. Not all reduction products formed at the disk reach the ring electrode and can be detected. A part of the products diffuses into the bulk solution instead. The fraction that is detected at the ring electrode is a number specific for the ring-disk electrode and is called theoretical collection efficiency (N 0 ). If I R /I D is normalized by N 0 , the transfer ratio is obtained.
At 0.66 V all species are reoxidized and I R /I D /N 0 should be equal to 1. The transfer ratios of the Cp 2 TiCl 2 solution are depicted in Figure S4.   (2). K 2 * in the base electrolyte amounts to 0.11 mM.   From the i D voltammograms three important observations can be made each of which will be addressed below. I) The limiting current density i lim at low potentials is decreasing in value with increasing concentration of L1. II) In the measurement with 1 mM of L1 (0.5 equivalents) a shoulder in the current step is formed. III) The new reduction potential found after addition of L1 is gradually shifting to more positive values with increasing concentration of L1.
The limiting current density at the disk electrode is related to the diffusion constant D of the analyte, the angular rotation frequency ω, the kinematic viscosity of the electrolyte ν and the bulk concentration of the analyte (Levich equation): [S10] !"#,! = −0.62 The other symbols have the usual meaning. The change of i lim is an indication for the formation of a new species with a smaller diffusion coefficient, which is the [Cp 2 Ti(IV)Cl 2 ]*L1 adduct. The gradual change of i lim can be explained with the reversibility of the adduct formation (3). On its path to the disk electrode surface [Cp 2 Ti(IV)Cl 2 ] can associate to L1 and dissociate again. The higher the concentration of L1, the larger is the diffusion length in the associated species. As a consequence, the

GC-Disk
III-c S17 diffusion of [Cp 2 Ti(IV)Cl 2 ]*L1 is best reflected in the measurement with 10 mM of L1. In Figure S7 the diffusion limited currents at the disk for Cp 2 TiCl 2 and with 10 mM and with 1 mM L1 (here the shoulder current was taken) are plotted against the square root of the rotation frequency. [Cp 2 Ti(IV)Cl 2 ]*L1 can be obtained using (6). This ratio amounts to 1.15: The diffusion coefficient D([Cp 2 Ti(IV)Cl 2 ]) for our system is not known in the literature. However, it can be easily calculated from the slope in Figure S7 (blue trace), if the kinematic viscosity ν is known.
Since ν scales to the power of -1/6 (see (6)) it is acceptable to approximate this value by using ν(THF) without conducting salt. From the literature a value of 5.2×10 -3 cm 2 s -1 is obtained. [S11] Thus, 8.8×10 -6 cm 2 s -1 results for D([Cp 2 Ti(IV)Cl 2 ]) and 7.6×10 -6 cm 2 s -1 for D([Cp 2 Ti(IV)Cl 2 ]*L1). The diffusion coefficient ratio together with the slopes of the base electrolyte and of the linear fit of the shoulder currents can be used to determine the equilibrium constant of association for (3) as the current -and as a result the slope m -in the shoulder is proportional to the concentration of the adduct in this equilibrium. The following relation is obtained: Here, c 0 denotes the initial concentration of Cp 2 TiCl 2 . K a then amounts to 0.69 mM -1 .

S18
The ring potential for the measurements with Cp 2 TiCl 2 and L1 was set to 0.05, 0.30 and 0.55 V vs.
Ag/Ag + . Analogous to before, the detection of [Cp 2 Ti(III)Cl 2 ] -, [Cp 2 Ti(III)Cl 2 ] -*L1 and [Cp 2 Ti(III)Cl] was achieved step-wise. The transfer ratios as a function of the concentration of L1 are given in Figure S8. The transfer ratios yield the ratios of the equilibrium concentrations for equations (4) and (5).
For the association of L1 to the anionic Ti(III) species K 1 amounts to 2 mM -1 and for the cleaving of Cl -*L1 off Cp 2 TiCl 2 -*L1 K 2 equals 2.8 mM.
The observed gradual shift in the reduction potential with increasing amount of L1 gives insight into the overall mechanistics of the equilibria in both oxidation states of Ti. It implies that the Nernst equation [S12] for this redox process depends on the concentration of free thiourea.
K' is introduced as a constant factor to account for the complex mesh of equilibria. A plot (see Figure S9) of the half-wave potential E 1/2 as a function of the decadic logarithm of c(L1) shows a slope of roughly 60 mV dec -1 at all rotation frequencies. Therefore, p in (11) must be equal to 1, which is indicative of 1:1 complexation between L1 and the titanocene species. With all experimental findings in mind we propose the electrochemical meshscheme shown in Scheme S1 for the Cp 2 TiCl 2 /L1 couple.
Scheme S1. Electrochemical mesh-scheme for the redox system of Cp 2 TiCl 2 in the presence of L1.
It considers the E q C r mechanism of titanocene dichloride and adds the L1 containing species. With the equilibrium constants we can quantify how effectively the C r reaction (cleaving off of 'Cl -') is amplified in the presence of L1 by comparing K 2 to K 2 *. This gives a factor of 25.

Sulfonamide L3
In contrast to thiourea L1 the sulfonamide L3 shows a double peak in the oxidation wave of 'Cp 2 TiCl'.
This lead us to assume the presence of an adduct between the active species and L3, [Cp 2 Ti(III)Cl]*L3.
Thus, the following equilibrium reactions have to be considered: RRDE measurements with Cp 2 TiCl 2 and L3 were performed in the same manner as described for L1.
The voltammograms of the 4 and 25 Hz measurements and the corresponding transfer ratios at 0.05, 0.23, 0.42 and 0.66 V ring potential are depicted in Figures S10 and S11.
IV-c Figure S12. Diffusion limited current densities at the disk i D as a function of f 1/2 for Cp 2 TiCl 2 and with 10 mM of L3 or 1 mM of L3 (shoulder current). The slopes of the linear fits are given as m.
By setting the ring potential for the measurements to 0.05, 0.23, 0.42 and 0.66 V, all Ti(III) species could be detected as in the previous cases. The transfer ratios as a function of the concentration of L3 are given in Figure S13.

S23
The transfer ratios yield the equilibrium constants for (13) to (15) by following the same procedure as for L1 (compare (9) and (10) A plot (see Figure S14) of the half-wave potentials E 1/2 of the disk voltammograms as a function of the decadic logarithm of c(L3) yields a slope of 60 mV dec -1 . Figure S14. Plot of the half-wave potential E 1/2 of the RRDE measurements with Cp 2 TiCl 2 and L3.
This implies a 1:1 complex formation for the Cp 2 TiCl 2 /L3 couple as for L1 before. We propose the electrochemical mesh-scheme shown in Scheme S2 for the Cp 2 TiCl 2 /L3 couple.
Scheme S2. Electrochemical mesh-scheme for the redox system of Cp 2 TiCl 2 in the presence of L3.
The mesh-scheme with L3 looks very similar to the one for L1. Only the occurrence of [Cp 2 Ti(III)Cl]*L3 makes an addition to the scheme necessary. According to the equilibrium constants K 2 and K 2 * the C r reaction is amplified by a factor of 27 in the presence of L3. This is comparable to L1. However, the access to [Cp 2 Ti(III)Cl]*L3 increases the extent of the C r reaction further.

Squaramide L2
Squaramide L2 similar to L3 exhibits a complicated oxidation wave in the range of the neutral Ti(III) with at least two oxidation peaks. Therefore we considered the same equilibrium reactions as we did for L3: The voltammograms of the 4 and 25 Hz measurements and the corresponding transfer ratios at 0.05, 0.23, 0.42 and 0.72 V ring potential are depicted in Figures S15 and S16.
By setting the ring potential for the measurements to 0.05, 0.23, 0.42 and 0.72 V, all Ti(III) species could be detected. The transfer ratios as a function of the concentration of L2 are given in Figure S18. The transfer ratios yield the equilibrium constants for (17)

S27
A plot (see Figure S19) of the half-wave potentials E 1/2 of the disk voltammograms yields a slope of 30 mV dec -1 for L2. Figure S16. Plot of the half-wave potential E 1/2 of the RRDE measurements with Cp 2 TiCl 2 and L2.
The smaller slope in the plot must originate from a different Nernst equation. Here the redox potential is proportional to log(c(L2) q ) wih q = p/2 (compare (11)). This can be explained with  (19), different values for K 1 ' (0.2 mM -1 ) and K 3 ' (0.1 mM -2 ) are obtained. K 2 remains unchanged. In reality, the equilibria from (19) and (20) are likely to be present side by side.
The given numbers for K 1 ( ' ) and K 3 ( ' ) can therefore only be considered as a simplification. We propose the electrochemical mesh-scheme shown in Scheme S3 for the Cp 2 TiCl 2 /L2 couple.

S28
Scheme S3. Electrochemical mesh-scheme for the redox system of Cp 2 TiCl 2 in the presence of L2.
The mesh-scheme with L2 looks very similar to the one for L3. However it adds the formation of the 2:1 adduct. According to the equilibrium constants the C r reaction is amplified by a factor of 491 in the presence of L2. The access to [Cp 2 Ti(III)Cl]*L2 and ([Cp 2 Ti(III)Cl]) 2 *L2 increases the extent of the C r reaction even more.

Comparison
In order to easily compare the solutions of Cp 2 TiCl 2 with L1 -L3 it is helpful to calculate the solution composition of the discussed titanocene species in %. Table S1 shows the compositions before and after reduction for a 1:1 mixture of 2 mM Cp 2 TiCl 2 and each additive. For L2, (20) is not considered here. Table S1. Solution compositions of titanocene species before (left) and after reduction (right) for 2 mM Cp 2 TiCl 2 and with the additives L1 -L3 in a 1:1 mixture in %. [

General Information
All visualizations of structures were created with UCSF Chimera [S13] 1.14.0.
We used the xTB, [S14] CREST [S15] and ENSO [S16] programs to determine the structures with the lowest free energy in THF solution for each ligand L and all complexes. These calculations were conducted in the same workflow.
Manually prepared starting structures, which are initially pre-optimized with the GFN2-xTB[GBSA] tight binding model, are used in the CREST program that employs MTD at the same level in order to obtain a relative complete ensemble of likely structures. The ENSO program determines the equilibrium (Boltzmann) populations for a few low-lying conformers at higher theoretical levels in three steps using TURBOMOLE. [S17] First, already relatively accurate B97-3c[DCOSMO-RS(THF)] (a composite low-cost DFT method [S18] ) single point energies are calculated on the CREST ensemble. Structures within an energy threshold of 4 kcal mol -1 above the lowest lying structure are then fully optimized at the same level. In this first filtering step thermostatistical free energies in the modified rigid-rotor/harmonic-oscillator (mRRHO) approximation [S19] calculated with GFN2-xTB[GBSA] and the free energy of solvation in THF calculated with the accurate COSMO-RS [S20] solvation model are added. Finally, for all structures within a 2 kcal mol -1 threshold an even better single point energy is computed at the PW6B95-D3/def2-TZVPP [S21] hybrid DFT level which basically replaces the corresponding B97-3c energy. In summary, the final complete total free energy used consists of the mRRHO part from the GFN2-xTB treatment, the COSMO-RS part in THF for solvation and the basic electronic energy with the PW6B95-D3 functional. In the following, the conformer of each species with the lowest total free energy is given, if not stated otherwise.

Calculated Structures
The basic receptor ability of the discussed additives can already give useful information for later analysis. The calculated structures of the H-bond donors L1, L2 and L3 are given in Figure S17. These conformers are the minimum structures in the electric field of a capacitor resembling the ε r of THF. Figure S17. Minimum structures of the H-bond donors L1, L2 and L3 obtained from DFT calculations. The energetic differences (ΔG 298.15 ) to the next higher lying conformers are given in kcal mol -1 . The number shown for L2 is the energy difference to the respective conformer in the first optimization step (E el ) as no conformer was within the threshold.
L1 shows a preferred orientation that allows H-bonding to a single H-bond acceptor with both N-H groups. In contrast, the N-H groups in L2 and L3 point into different directions. In order to properly discuss the reaction energies of the adduct formations, the binding affinity of the anion receptors to solvent molecules has to be assessed. The corresponding structure optimizations can be considered as a first approximation to explicit solvation. The direct anion receptor ability of each additive can be estimated by complexation with a simple anionic species. The Clanion is a natural choice in the context of our investigations. Therefore, expanding on the details given in the manuscript the supramolecular complexes of the receptors with THF and Clare depicted in Figure S18. The calculations of the [Cp 2 Ti(IV)Cl 2 ]*L2 adduct revealed a van der Waals complex as lowest energy structure. It does not align with the experimental observations but is depicted in Figure S19 for reference. Figure S19. Van  ΔG 298.15 = 6.9 kcal mol -1 ΔH 298.15 = 7.5 kcal mol -1 S32 aspects have been highlighted in the manuscript. In Figure S20 we show the full set of calculated structures.    Figure S21. In both structures coordination of a carbonyl oxygen of L2 to the Ti center can be observed. A calculation of ([Cp 2 Ti(III)Cl](THF)) 2 *L2 resultet in a van der Waals complex, which would have to be considered as a 1:1 complex in the context of our investigations.   Cl-*L1 (Fig. 9, entry 1; Fig. S18 Cl-*L3 (Fig. 9, entry 3, Fig. 18 Cp2TiCl2*L3 (Fig. 13, entry 3 Cp2TiCl2-*L2 (Fig. 10, entry 2 Cp2TiCl(THF)*L2 (Fig. 11, entry 2; Fig. S20 (Cp2TiCl)2(THF)*L2 ( Fig. 12; Fig. S21